Autonomous Trajectory Guidance under Uncertain Dynamical Environment Using State Transition Tensors

This study presents a robust autonomous guidance method using state transition tensors (STTs), which approxi-mate the dynamics through a higher-order Taylor expansion around a reference trajectory. The STTs approximation is applied to tube stochastic differential dynamic programming for the fast computation of the stochastic trajectory optimization. The rapid optimization using STTs enables redesigning a trajectory onboard after the uncertain physical parameters are estimated through in-situ observations along a reference tra-jectory. The proposed algorithm is numerically demonstrated with an orbital transfer scenario in the vicinity of a small body with uncertain gravitational parameters. The optimal policy can be obtained much faster, and it is equivalent to the original TSDDP method. This result shows the feasibility of autonomous guidance in an uncertain dynamical environment even with the constrained computational and fuel resources of spacecraft.